Self-generated prompt injections in compaction summaries | Hacker News - But Better

Self-generated prompt injections in compaction summaries simonwillison.net

In Our framework for reporting model misalignment OpenAI provide "six reports on unexpected or concerning model behavior we’ve observed in the last six months". This one here is my favorite: …

2 points•hn9zmdcaou•11 days ago•0 comments•

#generated #compaction #self #prompt #summaries #injections #self generated #generated prompt

0 comments

No comments yet.

Related stories

Self-Replicating Prompt Injections Turn Agent Context into an Open Relaydev.to
DEV Community · 3 points · 4 days ago
OpenAI: Self-replicating prompt injections existalignment.openai.com
Hacker News · 1 points · 4 days ago
Self-replicating prompt injections existalignment.openai.com
Hacker News · 1 points · 4 days ago
Show HN: Training a model to identify AI web content from structure alonearxiv.org
Hacker News · 73 points · 9 days ago
Show HN: Air-gapped file encryption as self-decrypting HTML pagecms-sfx-demo.apeleg.com
Hacker News · 90 points · 7 days ago
Prompt Wars – An AI riff on Core Wars to prompt-inject each otherpromptwars.party
Hacker News · 2 points · 10 days ago